P2 Sub-Batch 3 Ingest Report

Date: 2026-05-12 Batch: 50 papers listed in P2 sub-batch 3 task


1. Summary

CategoryCount
Papers in batch list50
Files present in raw/markdown3
Files missing from filesystem47
Source pages created3
False positives (out of scope)1 (FM_11123258)
Food concentration papers1 (FM_11125852)
Environmental/water papers1 (FM_11120698)

2. Classification Table

HandleCite KeyClassificationKey Finding
FM_11120698bousquet2024-lead-drinking-water-universityEnvironmental/water — Pb in drinking water8.43% of 5,954 fixture tests above 1 ppb LOD; max 1,100 ppb; UNC-CH campus; not food matrix
FM_11123258atanasov2024-sers-ammonium-nitrate-alnFalse positive — materials science SERSAl nanostructure SERS for ammonium nitrate; no food matrix; no food heavy metal data
FM_11125852cantoral2024-lead-levels-mexican-foodsFood concentration paper — reclassify P2 to P1Pb in 103 Mexican foods; infant rice cereal 1,005 ppb; soy infant formula 35 ppb; AAS; Mexico City retail
FM_11134948 through FM_12231472 (47 handles)various unknown2024-rsc-advances-*FILE MISSINGNot in raw/markdown; not in triage manifest

3. Food Concentration Paper Detail

FM_11125852 — cantoral2024-lead-levels-mexican-foods

Reclassification recommendation: P2 → P1. This is a primary food occurrence survey with direct concentration values in relevant infant food matrices. Meets all HMT&C Path A criteria.

Metal: Pb (total Pb; GF-AAS; no speciation) Jurisdiction: MX (Mexico City retail stores and markets) n: 103 food, beverage, and spice samples LOQ: 0.0025 mg/kg Method: Graphite furnace AAS (Perkin Elmer AAnalyst-600 with HGA-600); CRM bovine liver NIST 1577c; recovery 105.03 ± 9.01%; duplicate analysis per sample Basis: Wet weight for liquids; dry weight for solids dehydrated at 80°C for 72 h

Key concentration values (mg/kg = ppm; convert × 1000 for ppb):

Food itemPb mean (mg/kg)SDppb equivalentFAO/WHO ML (mg/kg)Exceeds ML
Infant rice cereal (Brand 2)1.0050.0421,0050.20Yes
Soy infant formula (Brand 2)0.0350.008350.01Yes
Whole wheat bread0.4470.1924470.20Yes
Pre-cooked rice0.2760.0172760.20Yes
Black pepper (bulk)0.2390.007239No ML
Turmeric (bulk)0.1760.032176No ML
Beef liver0.1330.0151330.20
Sweet bread0.1230.0201230.20
Paprika (bulk)0.0920.02792No ML
Wheat flour Brand 20.0700.007700.20
Pork ham0.0620.003620.15
Tamarind candy0.0500.001500.10 (FDA candy)
Guajillo chili0.0370.000370.05
Infant formula soy Brand 20.0350.008350.01Yes
Cacao powder (bulk)0.0830.032830.10 (FDA candy)
Soybean0.0290.013290.10
Turkey sausages Brand 20.0260.010260.10
Wheat flour Brand 10.0310.009310.20
Rice cake Brand 10.0300.012300.20
Tea sachet0.0210.00321No ML

All remaining 84 samples: <LOQ (<0.0025 mg/kg; <2.5 ppb)

Row-fit notes: Each food item is a single commercial sample (n=1 per item per brand). High within-item variance expected; SD for infant rice cereal is 0.042 mg/kg reflecting duplicate measurement variation, not between-sample variation. Row-fit is exact for matrices where the paper specifies (infant rice cereal, soy infant formula, pre-cooked rice, wheat flour, etc.). No speciation performed; values are total Pb.

Comparison context: The 1,005 ppb infant rice cereal value is the highest single-sample Pb concentration for that matrix in current wiki coverage. FDA Closer to Zero draft action level for processed foods eaten by babies/young children is 20 ppb as-consumed (fruit/vegetables) or may be referenced against powder-equivalent; the paper does not address reconstitution. This value was purchased retail in Mexico City; it is not representative of US market supply.


4. Testing Page Notes

FM_11123258 is a materials science paper on laser-structured AlN ceramic for SERS detection of ammonium nitrate. No food-relevant testing content. The Al and As flags in text-mining are from the chemical notation AlN (aluminium nitride) and As (arsenic) used in plasmonics substrate comparisons, not from food analysis. No testing page update warranted.


5. Reclassification Recommendations

FM_11125852: Reclassify P2 → P1. The manifest flagged this as a LOQ source candidate based on method validation language in the abstract. The paper is actually a multi-food Pb occurrence survey with primary concentration data across 13 food groups including infant rice cereal and infant formula. It meets all HMT&C Path A criteria.


6. Missing Files — Action Required

47 of 50 handles in this batch (FM_11134948 through FM_12231472) are absent from both raw/markdown/ and raw/manifest/triage-manifest.csv. The triage manifest has 23,260 rows and was built 2026-05-12; these handles are not in it.

Git status shows two untracked paths: raw 2.zip and raw 2/. These are likely the source of the missing Marker-converted markdown. The handles in the batch list follow the manifest numbering scheme above the current highest ingested range, suggesting they belong to a not-yet-imported batch.

Action needed (Karen): Confirm whether raw 2/ or raw 2.zip contains the markdown for these handles. If yes, the workflow is:

  1. Copy the FM_111xxxxx through FM_122xxxxx folders from raw 2/ into raw/markdown/
  2. Extend the triage manifest with rows for the new handles
  3. Re-run this sub-batch against the updated manifest

Until the files are available in raw/markdown/ with manifest entries, no source pages can be created for these 47 handles. They are logged here as missing, not silently dropped.


7. Source Pages Created

  1. wiki/sources/cantoral2024-lead-levels-mexican-foods.md — Food concentration paper; Pb in 103 Mexican foods; P1 reclassification recommended
  2. wiki/sources/bousquet2024-lead-drinking-water-university.md — Pb in drinking water at UNC-CH; environmental exposure paper; not food matrix
  3. wiki/sources/atanasov2024-sers-ammonium-nitrate-aln.md — Materials science / SERS false positive; no food data